hyper-sinh: An accurate and reliable function from shallow to deep learning in TensorFlow and Keras

This paper presents the ‘hyper-sinh’, a variation of m-arcsinh activation function suit-able for Deep Learning (DL)-based algorithms supervised learning, including Convolutional Neural Networks (CNN) and Recurrent (RNN), such as Long Short-Term Memory (LSTM). hyper-sinh, developed in open-source Python libraries TensorFlow Keras, is thus described validated an accurate reliable shallow deep neural networks. Improvements accuracy reliability image text classification tasks on six (N=6) medium-to-large benchmark datasets are discussed. Experimental results demonstrate that overall competitive performance novel hyper-sinh networks yielded highest performance. Furthermore, this evaluated against other gold standard functions, demonstrating its both tasks.